Systems methods and articles of manufacture for analyzing on-line banking account data using hybrid edit distance

1. A computer-implemented method for selecting and performing optimized processing for determining whether an on-line banking account is utilized for business purposes based on a meaning associated with an abbreviation in an electronic description of the on-line banking account, the method comprising: receiving or determining, by an intermediate computer in communication with a computer of a financial institution and a computer of an account holder through respective networks, the electronic description of the on-line banking account the account holder has at the financial institution, the electronic description comprising the abbreviation; accessing, by a rule processor comprising computer-executable instructions executed by a processor of the intermediate computer, a data store comprising a set of rules and a plurality of pre-determined normalized strings; applying, by the rule processor, the set of rules to the abbreviation and the plurality of pre-determined normalized strings to select processing for determining respective edit distances representing respective degrees of transformation from the abbreviation to respective pre-determined normalized strings, the set of rules comprising: a first rule specifying that when a first letter of the abbreviation is the same as the first letter of the normalized string for a pair comprising the abbreviation and a normalized string, then first processing for determining an edit distance for the pair is to be applied, the first processing determining the edit distance without considering any vowel appearing after a first letter of the normalized string, a second rule specifying that when the first letter of the abbreviation is the same as the first letter of the normalized string, then second processing for determining the edit distance for that pair is to be applied by the intermediate computer applying at least one of a Levenshtein distance, a Damerau-Levenshtein distance, and a Hamming distance to all consonants of the normalized string following the first letter, and a third rule specifying that when the first letter of the abbreviation is not the same as the first letter of the normalized string, then third processing for determining the edit distance is to be applied by the intermediate computer applying at least one of the Levenshtein distance, the Damerau-Levenshtein distance, and the Hamming distance to all letters of the normalized string, wherein, for each respective pre-determined normalized string: the first processing is selected in response to the first rule being satisfied, the second processing is selected in response to the second rule being satisfied, and the third rule is selected in response to the third rule being satisfied; performing, by the intermediate computer, the selected processing from among the first processing, second processing, and third processing to determine the respective edit distances representing the respective degrees of transformation from the abbreviation to the respective pre-determined normalized strings, wherein: a first edit distance generated by performing the selected processing on a pair comprising the abbreviation and the normalized string selected as corresponding to the abbreviation is less than a second edit distance generated by applying the Levenshtein Distance to the pair comprising the abbreviation and the normalized string corresponding to the abbreviation, thereby more clearly indicating which normalized string of the plurality of normalized strings should be selected by the rule processor as corresponding to the abbreviation, a first difference between the first edit distance and a third edit distance, determined by the rule processor applying the set of rules to a second pair comprising the abbreviation and a normalized string that does not correspond to the abbreviation and is larger than the first edit distance, is larger than a second difference between a fourth edit distance and a fifth edit distance determined by the Levenshtein distance, thereby more clearly indicating which normalized string of the plurality of normalized strings should be selected by the rule processor as corresponding to the abbreviation, wherein the fourth edit distance is determined by the Levenshtein distance to the pair comprising the abbreviation and the normalized string selected as corresponding to the abbreviation, and the fifth edit distance is determined by execution of the Levenshtein algorithm to the second pair comprising the abbreviation and a normalized string that does not correspond to the abbreviation and is larger than the third edit distance; selecting, by the rule processor, a normalized string from the plurality of pre-determined normalized strings as corresponding to the abbreviation based upon the determined respective edit distances for respective pairs of the abbreviation and respective normalized strings as obtained by the performing; automatically determining, by the rule processor, whether the on-line banking account is utilized for business purposes based upon the selected normalized string indicating that the meaning associated with the abbreviation is business-related; and in response to determining that the on-line banking account is utilized for business purposes, displaying, by the rule processor, a message related to the on-line banking account through a display of the computer of the account holder.

2. The method of claim 1 , the intermediate computer receiving the electronic description through a network from the computer of the financial institution.

3. The method of claim 1 , the electronic description comprising a description of the on-line banking account that was entered or selected by the account holder.

4. The method of claim 1 , the plurality of pre-determined normalized strings forming a specialized or domain-specific reference dictionary that is a subset of terms of a general dictionary and accessed by the rule processor.

5. The method of claim 4 , further comprising: generating, by the intermediate computer, the specialized or domain-specific reference dictionary; and storing, by the intermediate computer, the specialized or domain-specific reference dictionary to the data store.

6. The method of claim 1 , further comprising: determining, by the intermediate computer, respective edit distances for respective pairs of the abbreviation and respective normalized strings; and selecting, by the rule processor, the normalized string of the pair having the lowest edit distance.

7. The method of claim 1 , further comprising selecting, by the rule processor, a normalized string of a pair having an edit distance that is not an edit distance determined by the Levenshtein distance Hamming distance or the Damerau-Levenshtein distance.

8. The method of claim 1 , wherein multiple account holders have respective on-line banking accounts at the financial institution, the method further comprising applying, by the rule processor, the set of rules to respective descriptions of respective multiple on-line banking accounts of respective account holders in order to determine which of the respective on-line banking accounts are business accounts or utilized for business purposes.

9. The method of claim 1 , further comprising: automatically determining, by the rule processor, that the on-line banking account is an account of a business; and presenting, by the intermediate computer, wherein the message comprises an offering of a good or service to the account holder, the offering being related to the business.

10. The method of claim 1 , further comprising: automatically determining, by the rule processor, that the on-line banking account is an account of a business; and presenting, by the intermediate computer, wherein the message comprises an incentive to the account holder to remain at the financial institution.

11. The method of claim 1 , wherein the intermediate computer receives the description from the account holder when the on-line banking account is opened, wherein the description does not identify the account holder as a business or indicate that the account is utilized for business purposes.

12. The method of claim 1 , wherein the intermediate computer receives the description from the account holder when the account holder opened the on-line banking account with the financial institution.

13. The method of claim 1 , the on-line banking account including at least one of a savings account, a checking account, a money market account, and a credit card account.

14. The method of claim 1 , further comprising: receiving or determining, by the intermediate computer electronic transaction history data of the on-line banking account, the electronic transaction history data including an abbreviation; applying, by the rule processor, the set of rules to the abbreviation of the electronic transaction history data and a plurality of pre-determined normalized strings to determine respective edit distances representing respective degrees of transformation from the abbreviation of the electronic transaction history data to respective pre-determined normalized strings; and selecting, by the rule processor, a normalized string from the plurality of pre-determined normalized strings that corresponds to the abbreviation of the electronic transaction history data based at least in part upon the determined respective edit distances for respective pairs of the abbreviation of the electronic transaction history data and respective normalized strings.

15. The method of claim 1 , wherein application of the first rule to a pair of the abbreviation and a normalized string by the intermediate computer results in generation of a numerical result that is less than a numerical result generated by application of the Levenshtein distance to the same pair.